Adding a camera to a subsequence through scripting creates the binding in the incorrect binding space

Sequencer scripting creates all binding IDs in the root binding space, so when the bindings exist in sub sequences and shots, they will not resolve unless those are opened through the sequence that contains them. The linked UDN provides a python script that reproduces the issue. 

1. Create a level with a camera in it
2. Use python scripting to:
   1. Create a level sequence (master sequence)
   2. Add a shot track to that sequence
   3. Add a section to the shot track
   4. create a new level sequence (shot)
   5. Set the sequence on the shot track section to the newly made shot
   6. Add a camera cut track to the shot
   7. Add a section to the camera cut track
   8. Add the camera in the level as a possessable to the shot 
   9. Create a binding ID for the camera possessable by calling make_binding_id on the master sequence
   10. Set the camera binding ID on the camera cut section to the binding ID from #2.9
3. Open the newly created master sequence, and double click on the shot section to open the shot, and lock to the camera cut track in the shot. The viewport should lock correctly
4. Browse to the shot asset and open it on its own
5. Lock to the camera cut track again

RESULT
The viewport does not lock correctly, and the camera binding is unresolved.